Non-contact thermometers that absorb the heat radiating from your skin can help to check masses of people for fever, but the numerous proofs of their accuracy are unsatisfactory.
Your forehead is about to get a lot of action. Is it greasy? You may want to wipe off the sweat. In the age of COVID, our forehead becomes the target of fever samples. Thermal imaging cameras and non-contact infrared thermometers are used to prevent people with fever from entering an area where they can infect others. The question is only whether these temperature detection technologies actually work.
Thermometers date back to the early 1600s, but even before they existed, doctors recognized that there was heat and cold. Johannis Hasler from Bern, a theologian and physician, believed that body temperature had risen in tropical latitudes and published a table of presumed body temperatures around the world to determine his mixture of “medicines”. Thermometers were originally based on the expansion of water as it warmed up; this water was later replaced by mercury, which, because of its mercury toxicity, reacted much faster to temperature changes than alcohol. More recently, liquid in electronic thermometers has been abolished, with many working on the principle that electrical resistance changes with temperature. And now contact with the body is no longer necessary. There is a class of devices called thermal imaging cameras, which includes both thermal imaging cameras at some airports and non-contact infrared thermometers (NCITs). All of these thermal imagers operate on the principle that heat is emitted from any object above absolute zero (-273.15°C or -459.67°F). Our forehead emits heat in the form of infrared radiation. An NCIT has a lens that focuses this heat onto a detector that converts it into electricity. And because the temperature of the forehead is lower than the temperature under our tongue, the device must convert it into an “oral temperature equivalent”. The laser you see on some devices does not detect temperature. It is there to ensure that the user aligns the device correctly.
Thermal imaging cameras have their advantages, especially for mass screening of people during a pandemic. They are relatively easy to use. The lack of contact should reduce the spread of the disease. And they quickly pick up a temperature and resume it.
But if we concentrate on forehead infrared thermometers – sometimes called “thermometer guns” – for one minute, theoretically a lot can affect their accuracy. They should not be used in direct sunlight, as the sun warms your forehead and affects the reading. Sweat on your forehead can artificially lower the measured temperature and essentially mask fever. There is actually a significant list of reasons why these infrared thermometers can miss fevers: circulatory problems, previous injuries, heavy makeup, certain medications. On the other hand, the device can tell you that you have a fever even though you are not suffering from stress, physical activity, nicotine, a hot drink, hormonal treatments, pregnancy, hot air currents or even sunburn. And we can easily imagine a scenario where someone with COVID is contagious and takes anti-fever medication such as paracetamol, which lowers the temperature.
These are the built-in pitfalls of forehead infrared thermometers. How do they compare in real life to other types of thermometers?
We need to cool down our expectations
The gold standard for assessing body temperature is not the rectal temperature, but the core temperature. This may mean inserting a catheter into the artery in a person’s lung to measure the temperature of their blood, as the brain senses this when it adjusts the temperature of your body. Suffice it to say that in a pandemic, doctors cannot avoid inserting catheters into people’s pulmonary arteries to check for fever. Rectal temperature is a decent indicator, but its status as such has been criticised by doctors as essentially undeserved. However, a recent meta-analysis of studies comparing a variety of thermometers to the core temperature (the gold standard) found that rectal and oral electronic thermometers were the best, with in-ear thermometers (commonly known as eardrum thermometers) being the last to die.
Non-contact infrared thermometers, which include the eardrum and the forehead thermometer, have been studied, and the limited knowledge of these devices is a microcosm of the problems that affect scientific clarity. There is a study, but they have only tested one model. How representative are their results? Here there is a study, but it was carried out on infants by trained health professionals: Are the results in an adult population examined by non-medical personnel? There is an overview of many studies, which highlights the fact that the studies may have been biased. The point is that the accuracy of these infrared thermometers is difficult to determine because the studies are deeply incomplete. I am tempted to give the figures that emerged from a meta-analysis in 2009. However, since NCITs have been compared to the eardrum temperature and the eardrum temperature is not reliable, I am not sure we can trust these figures.
Finally, an additional wrinkle in the armor: a group of English scientists tested nine common NCITs for a range of measures and found that five of them were outside the accuracy range specified by the manufacturer and the medical standard they were supposed to meet.
For thermal imaging cameras (many of which skip the forehead and measure the temperature in the inner corner of the eye), a recent summary of the evidence published by The Conversation pointed to a similar lack of evidence, while raising the specter of privacy issues. These cameras are often able to identify faces, and the authors wondered if the facilities that collect this personal information are storing it securely.
What should we do with it? A team from the Food and Drug Administration has argued that infrared thermometers are the only viable way to screen the public for infectious diseases. However, there is not a single temperature threshold that everyone agrees defines as a fever, because the normal body temperature is within a certain range. Also, you do not need COVID-19 to get a fever, and you do not need a fever to get COVID-19. And due to the lack of good data on the accuracy of infrared thermometers, it is easy to ask on what basis we can justify their use. The below-average studies may convince us that these devices can effectively rule out fever, but these studies mainly concern health care workers who take the temperature of babies. What happens if poorly trained people who are concerned about coronavirus use these devices far away from the 3 to 15 cm recommended by the manufacturer?
The question that remains is this: Thermal imaging devices such as non-contact infrared thermometers and thermal imaging cameras, while not very good, are good enough? Will their use be a net good in the middle of a pandemic? Unfortunately, the evidence base is mother on this issue. We need better studies.